Electrochemical studies of antimony (III) and antimony (V) in molten

In Situ STM Studies on the Underpotential Deposition of Antimony on Au(111) and Au(100) in a BMIBF4 Ionic Liquid. Yong-Chun Fu, Jia-Wei Yan, Yu Wang, ...
0 downloads 0 Views 909KB Size
Inorg. Chem. 1984,23, 1726-1734

1726

7721-01-9; NbCl,, 10026-12-7; D, 7782-39-0. SupplementaryMaterial AvallaMe: Listings of phosphine hydrogen atom positions, anisotropic thermal parameters and observed and calculated structure factors for TaC12H2(PMe3)4 (sa) and TaC12H2(dmpe)2( 6 4 and experimental ESR spectra of 6a and 6b (22 pages). Ordering information is given on any current masthead page. Complete structural reports on 5s (MSC Report No. 82934) and 6a (MSC Report No. 83913) are available, in microfiche form only, from the Chemistry Library, Indiana University.

Crystal data are given in Table V.

Acknowledgmqk The authors are grateful to the National (Grant CHE 82science Foundation for financial 06 169) and the Wrubel computing Center, Indiana University, for a generous gift of computing time. Registry No. 1, 85923-35-9; 2, 89656-65-5; 3, 61916-35-6; 4, 89708-59-8; Sa, 85939-38-4; 6a, 89708-60-1; 7a, 89708-61-2; T a C l , ( d m ~ e ) ~61916-34-5; , NbC14(dmpe)2, 61 202-65-1; TaCl,,

Contribution from the Department of Chemistry, State University of New York at Buffalo, Buffalo, New York 14214

Electrochemical Studies of Antimony(II1) and Antimony (V) in Molten Mixtures of Aluminum Chloride and Butylpyridinium Chloride DHIA A. HABBOUSH' and ROBERT A. OSTERYOUNG*

Received May 24, 1983 Electrochemical studies of Sb, Sb(III), and Sb(V) have been camed out in molten mixtures of AlC13 and N-1-butylpyridinium chloride (BuPyCl) at 40 OC, as a function of melt composition. Analysis of measurements in the acidic melts indicates SbC12+as the dominant species. The reduction of this species on glassy carbon exhibits irreversible behavior. In the basic melts, SbC14- and SbC1,- are believed to be the dominant species. The reduction of Sb(II1) to S b on glassy carbon also showed irreversible behavior while its oxidation to Sb(W revealed a quasi-reversible behavior. No Sb(II1) oxidation was observed in the acidic melts.

Introduction Molten mixtures of AlCl, and N-1-butylpyridinium chloride (BuFyCl) have been shown to be useful and interesting solvents for a variety of electrochemical and spectroscopic studies.'-7 The mixtures, of composition ranging from 2:l to 0.75:l (mol) of A1Cl3:BuPyC1, are liquids at essentially ambient temperatures2P8 They are ionic liquids, and their composition can be adjusted to change their acid-base properties which, in turn, determine their oxidation-reduction and coordination chemistry. Raman and infrared spectroscopic studies showed that the dominant Al-containing species in the system are AlC14and AlzC17- ion^.^.^ The acid-base equilibrium in these solvents has been studied potentiometrically, by employing A1 indicator electrodes, and the equilibrium 2AlC14- AlzCl7- + C1(1) was shown to describe the system across the entire range of compositions mentioned above.lOJ Electrochemical studies of several organic and inorganic solute species in this and similar systems revealed a pronounced dependence of the solute chemistry and electrochemistry on the system a ~ i d i t y . ~ In q~-~ the basic (excess BuPyCl) melt, stable chloro complexes of Ni(II), Fe(II), Fe(III), Cu(I), Cu(II), and Co(I1) have been reported.

396~12~1

Studies of molten SbC1, and its mixtures with alkali-metal chloride and/or AlCl, have reported the formation of several species of As a result of conductance measurements of AlCl, in molten SbC1314and of KCl in molten Sbc13,15,16 the presence of SbC12+and SEI, in these mixtures has been suggested, and the reaction SbCl, AlCl, SbClZ++ AlC14(2) was proposed for AlCl, in SbC13.14 Raman studies, however, did not give support for the above suggested species and, instead, indicated some interaction between SbCl, and AlCl,

+

-

f Present address: Department of Chemistry, Sacred Heart University, Bridgeport, CT 06606.

in their molten equimolar mixture and the possible formation of some higher chloro complexes of antimony in molten mixtures of SbC1, and KCl.17 SbCl2+,produced from reaction 2,has recently been postulated as the species responsible for the chemical oxidation of perylene in molten SbCl, containing AlC13.18 Lower valent Sb species have been suggested as a result of solubility measurements of Sb in molten SbCl, containing A1C1319,20 and CsC1.20 We have studied the electrochemistry of Sb in molten mixtures of AlCl, and BuPyCl to characterize the major species present. Such information would also be useful in any study of organic solutes in solutions of melts containing SbC1, where the solvent acidity can be ~ a r i e d . ~ * q ~ ~ - ~ ~

(1) Chum, H. L.;Koch, L. L.; Osteryoung, R. A. J. Am. Chem. Soc. 1975, 97, 3264. (2) Robinson, J.; Osteryoung, R. A. J. Am. Chem. SOC.1979, 101, 323. (3) Gale, R. J.; Gilbert, B.; Osteryoung, R. A. Znorg. Chem. 1979, 18,2723. (4) Robinson, J.; Osteryoung, R. A. J. Am. Chem. SOC.1980, 102, 4415. ( 5 ) Gale, R. J.; Osteryoung, R. A. J . Electrochem. SOC.1980, 127, 2167. (6) Nanjundiah, C.; Shimizu, K.; Osteryoung, R. A. J . Electrochem. SOC. 1982, 129, 2474. Nanjundiah, C.; Osteryoung, R. A. Ibid. 1983, 130, 1312. (7) Gale, R. J.; Job, R. Znorg. Chem. 1981, 20, 40, 42. (8) Gale, J. R.; Gilbert, B.; Osteryoung, R. A. Inorg. Chem. 1978,17,2728. (9) Gale, R. J.; Osteryoung, R. A. Inorg. Chem. 1980, 19, 2240. (10) Gale, R. J.; Osteryoung, R. A. Inorg. Chem. 1979, 18, 1603. (11) Schoebrechts, J.; Gilbert, B. J. Electrochem. SOC.1981, 128, 2679. (12) Hussey, C. L.; Laher, T. M. Znorg. Chem. 1981, 20, 4201. (13) Laher, T. M.; Hussey, C. L. Inorg. Chem. 1982, 21, 4079. (14) Texier, P. Bull. SOC.Chim. Fr. 1968, 4716. (15) Porter, G. B.; Baughan, E. C. J . Chem. Soc. 1958, 744. (16) Davis, A. G.; Baughan, E. C. J . Chem. SOC.1961, 1711. (17) Fung, K. W.;Begun, G. M.; Mamantov, G. Inorg. Chem. 1973, 22, 53. (18) Sorlie, M.; Smith, G. P.; Norvell, V. E.; Mamantov, G.; Klatt, L. N. J . Electrochem. Soc. 1981, 128, 333. (19) Corbett, J. D. "Progress in Inorganic Chemistry"; Lippard, S. J., Ed.; Wiley: New York, 1976; Vol. 21, pp 129-158. (20) Sorlie, M.; Smith, G. P. J. Inorg. Nucl. Chem. 1981, 43, 931. (21) Buchanan, A. C., 111; Dworkin, A. S.;Smith, G. P. J . Am. Chem. SOC. 1980, 102, 5262.

0020-166918411323-1726%01.50/0 0 1984 American Chemical Society

Inorganic Chemistry, Vol. 23, No. 12, 1984 1727

Electrochemical Studies of Sb(II1) and Sb(V) 3

I ' 0.8

E, V vs

1

ZOOPA

&

1.5

1.0

0.5 0 -0.5 E , V vs AI/AI(III)

0.6

-1.0

Figure 1. Cyclic voltammograms of Sb(II1) melt solutions: a, 7.0 X M in 1 3 1 melt, v = 20 mV s-I; b, 6.3 X M in 0.8:l melt, u = 5 mV s-'.

Experimental Section Preparation of AIC13 and n-butylpyridinium chloride has been described elsewhere.2 Sb(II1) was generated in situ by anodic dissolution of an antimony rod (Alfa Products) by constant-potential coulometry or by the addition of anhydrous SbCI3 (Alfa Products). The antimony rod was cleaned in concentrated HCI, washed with water and acetone, and dried. Aluminum wire (Alfa Products) was cleaned in a 30:30:40 volume mixture of concentrated sulfuric, nitric, and phosphoric acids, washed with water and acetone, and dried. A homemade glass container was employed as the electrolytic cell and was covered with a Teflon lid with holes for reference electrode and counterelectrodecompartments, a thermocouple well, an antimony rod, and the rotating disk electrode. The cell assembly was placed in a furnace whose temperature was controlled at 40 1 O C by a Thermo Electric temperature controller. A rotating glassy-carbon disk electrode (RGCDE), with an area of 0.196 cm2,was employed; its surface was polished with 0.2-pm alumina on a wet polishing cloth and washed in water. The reference electrode and counterelectrode were AI wires dipped in 2:l AIC13:BuPyCI melt, and both were separated from the rotating disk electrode compartment by fine glass frits (porosity E). An EG&G PARC 175 programmer and PARC 173 potentiostat with a Model 179 coulometer were used for all voltammetric and coulometric experiments, which were recorded on an HP 7046 X-Y recorder. Potentiometric measurements were made with a Keithly 168 digital voltmeter with a voltage follower in the circuit. A Pine Instrument Co. ASR rotator was employed for all rotating electrode experiments, which were camed out in a Vacuum Atmospheres drybox with a circulating, purified argon atmosphere. Melts were prepared by mixing weighed amounts of purified AlC13 and BuPyC1.*JOThe electrode surface was conditioned by holding at an initial potential, where no reaction of primary interest took place, until the background current had decayed to a low, constant value. Constant-potential coulometric experiments were performed by changing the potential from a value where no current passed to one

*

(22) Buchanan, A. C., III; Dworkin, A.S.;Smith, G. P. J. Org. Chem. 1981, 46, 47 1. (23) Buchanan, A. C., 111; Livingston, R.;Dworkin, A. S.;Smith, G. P. J . Phys. Chem. 1980,84, 423. (24) Buchanan, A. C., 111; Dworkin, A. S.;Brynestad, J.; Gilpatrick, L.0.; Poutsma, M. L.; Smith, G. P. J . Am. Chem. SOC.1979, 101, 5430.

0.4 AI/AI(III)

0.2

Figure 2. Voltammograms of a 6.2 X

M solution of Sb(II1) in 1.24:1 melt, at RGCDE. Rotation rate/rpm: a, 450; b, 650; c, 960; d, 1270; e, 1670. where the desired reaction would take place. The solutions were stirred during the coulometric experiments. Results and Discussion SbC13 dissolved readily in molten mixtures of AlC1, and BuPyCl (both acidic and basic) to form almost colorless solutions. Representative cyclic voltammograms of these solutions, at a glassy-carbon electrode vs. an Al/Al(III) reference, are shown in Figure 1 and yield one cathodic wave for solutions in acidic melts and two waves, one cathodic and one anodic, for solutions in basic melts. Solutions obtained from anodic dissolution of antimony metal at constant potentials in these melts (at +1.3 V in the acidic 2:l and at -0.2 V in the basic 0.75:l AlC13:BuPyC1melt) gave cyclic voltammograms similar to those of the SbC13 solutions. Acidic Melts. Table I lists data obtained from cyclic voltammetry of solutions of SbC13in acidic melts. The cathodic current peak for S b deposition shifts negatively with increasing scan rate, which is indicative of either slow charge-transfer reaction rate (irreversible behavior) or, more likely, nucleation overvoltage. A negative shift of this peak with decreasing acidity is also evident and would be expected if the complexing ligand is either one of two species present in these melts, namely C1- or AlCl.,-, rather than A12C17-. The reverse scan produced one stripping peak; the charge under the anodic stripping peak equaled that on the cathodic scan. Voltammograms of these solutions were also obtained on the RGCDE and are shown in Figure 2. Plots of the limiting cathodic currents at potentials on the plateau (ilc), for solutions in various acidic melts, vs. the square root of the rotation rate ( w ' / * ) were linear (Figure 3) and passed through the origin, indicating that the reaction is convective diffusion controlled in the limiting region. To confirm the oxidation state of the antimony species present in these solutions, constant-potential coulometry for the deposition of antimony (from 7.7 mL of a 11.1 X M SbCI3 solution in a 2:l melt) was carried out at +0.4 V on a glassy-carbon electrode of a large area; the +0.4 V value corresponds to a potential on the diffusion limiting plateau of the reduction wave (Figure 2). This electrolysis was stopped at various times, and voltammograms were run at a RGCDE. A plot of the cathodic charge (Q,) vs. i,, was linear, and thus Q, for a complete reduction to S b was obtained by extrapolation to the Q,axis. The results of this experiment established that the oxidation state of antimony in solution was +3; Qc found was 25 C compared with that calculated for a complete

1728 Inorganic Chemistry, Vol. 23, No. 12, 1984

Habboush and Osteryoung

Table I. Dependence of E P C E,,, , E"', and E"', for Sb(III)/Sb on Melt Composition NAINB

2.00

1.95 1.92 1.83 1.64 1.so 1.44 1.24 1.09 1.06 1.04 1.02 1.01 0.98 0.94 0.91

u, mV s"

10 20 50 100 200 5 00

20 20 20

20

0.665 0.640 0.580 0.550 0.5 15 0.455

0.530 0.495 0.435

0.400

EP,'

v

Ea',O'C

1.145 1.165 1.210 1.265 1.325 1.390

v

5 10 20

5 5

-0.775 -0.820 -0.865 -0.925 -0.885 -0.910

P',,b*c

1.007

0.380

0.977 0.963 0.938 0.90 1

0.399 0.402 0.403 0.397

0.877 0.848

0.396 0.390

v

1.110 1.080 1.045

0.81 1 0.773 0.758 0.736

1.015

-0.431 -0.47 1

so 0.87 0.81 0.80 0.75

EPc, v

0.387 0.362 0.368 0.392 av 0.389 ?: 0.014 -0.521 -0.521

EO',a,d

v

Eo',,b*d

0.988

0.359

0.833

0.346

0.755

0.323

0.708

0.331

v

-0.345 -0.320 -0.305 -0.280 -0.420 -0.430

-0.504 -0.516

-0.52s -0.523

-0.527

-0.523 av -0.523

?1

0.002

a E"' = E,, (RTI3F) In [ Sb(III)]. = Eeq - (RT/3F) In [ Sb(III)] + @R T/3F) In [ C1- J ;p = 2 for N A / N B > 1, p = 4 for N A / N B < 1. Prom potentiometry. From voltammograms of Sb(II1) solutions on RGCDE obtained from a reverse scan.

Table 11. Dependence O f &b(III) and qDSb(III) on the Melt Composition in the Acidic Region

2.00 1.50 1.24 1.09 1.01 a

0.145 0.175 0.195 0.20s 0.21s

8.15 6.88 6.02 5.76 5.58

1.18 1.20 1.17 1.18 1.20 av 1.19 k 0.01

Reference 6.

reduction of Sb(II1) to Sb, 24.3 C. The values of the limiting currents, obtained from the voltammograms at the RGCDE (Figure 2), were used to calculate the diffusion coefficients (D), of the Sb(II1) species in various acidic melts, from the Levich equation

i, = 0.620nFAD2/3w'/2u-'/3C* (3) where u is the kinematic viscosity (viscosity divided by density), c* is the concentration of the electroactive species in the bulk in units of mol ~ m - and ~ , the other symbols have their usual meanings. The results of these calculations, along with the products of vD,are tabulated in Table 11. The vD product was found to be constant across the entire range of acidity investigated for the Sb(II1) species in acidic melts. Thus, the Stokes-Einstein relation

D = kT/6vnr

(4)

where k is the Boltzmann constant and r is the radius of the diffusing species, was obeyed. Initial attempts to measure a stable value for the potential of the Sb(III)/Sb couple were unsuccessful. This variation in potential was attributed to a slight oxidation of the antimony metal into the melt. This oxidation was evident when cyclic voltammetry of a 2: 1 melt, following immersion of an antimony rod into it for 1 h, gave rise to a voltammogram similar to that

of curve a in Figure 1. The concentration of Sb(II1) after this period was estimated as 0.1 X lom3M. An additional 24-h immersion increased the concentration of the Sb(II1) by about 30%. The possibility that this formation of Sb(II1) resulted from dissolution of an oxide film on the Sb metal was ruled out when it was found that an antimony rod, previously held in a similar melt for several weeks, gave rise to a similar solubility of Sb(II1) after 18 h of immersion in a new melt. These observations are analogous to an apparent spontaneous oxidation of ferrocene by the acidic melt.25 Recent work on the reduction of "water" in the melt indicates that the reduction of "proton" takes place at potentials close to, but slightly positive of, the Sb-Sb(II1) potential in either acidic or basic melts.26 This suggests that Sb might have been oxidized by a proton-containing species, possibly an -Al-O-H entity,27to yield Sb(II1). Part of the reason may also have been a result of leakage of the melt from the bulk solution to the reference compartment, thus changing melt composition in that compartment, and hence the reference potential. While potentiometric measurements were ultimately obtained (see below), the following procedure was employed to obtain estimates of the Sb-Sb(II1) potential under the conditions referred to above. Voltammograms of Sb(II1) solutions on a RGCDE (Figure 4) obtained on a reverse scan, i.e., following oxidation of the deposited Sb metal from the electrode surface, gave a constant value for the zero crossing potential as a function of electrode rotation rate. This zerocurrent crossing potential, as a function of melt composition, was utilized as the equilibrium potential of the Sb(III)/Sb couple to obtain an indication of the nature and number of ligands bound to Sb(II1). Values of Eo' so obtained are shown in Table I for comparison with those obtained from direct ( 2 5 ) Karpinski, Z.; Nanjundiah, C.; Osteryoung, R.A., to be submitted for

publication.

( 2 6 ) Sahami, S.; Osteryoung, R. A. Anal. Chem. 1983,55, 1970. (27) Tait, S.; Osteryoung, R. A,, to be submitted for publication.

Inorganic Chemistry, Vol. 23, No. 12, 1984 1729

Electrochemical Studies of Sb(II1) and Sb(V)

7

6

5

4 t

4

I

$

--e

3

1

0

0.5 w1/2,

1.0

1*5

2.0

,,c-1/2

Figure 3. Limiting cathodic currents for the reduction of Sb(II1) vs. M in 2:l melt; 0, 7.0 X M in 1 3 1 melt; M in 1.09:lmelt; 0, M in 1.24:lmelt; 0 , 5.7 X 8 , 6.2 X 5.5 X lo-’ M in 1.01:l melt.

dZ: @, 8.2 X

I

-20

where E , is the Sb(III)/Sb equilibrium potential, Eo1is the formal potential of the couple in the absence of free ligand (Le., no complex formation), C,is the total concentration of all Sb(II1) species, [L-] is the concentration of the free ligand, CS is the standard concentration defined as 1 mol L-l, and the p’s are the overall stability constants for the various complex species. Equation 5 may be used to determine the number of ligands and to evaluate the stability constants. In order to do so we introduced a new formal potential Eo‘ defined as

Eo’= E o 1 - ( R T / n F ) In (1

+ p,[L-] + P2[L-I2 + ...)

(6)

which is independent of species concentration and can be calculated from the relation E,, = Eo’

10

+ ( R T / n F ) In (C,/Ce)

(7)

which is a reduced form of eq 5 . A plot of Eo’vs. In [Cl-] (Figure 5, curve a) gave a straight line over the melt composition range 2:l to 1.Ol:l, indicating one dominant complex. Hence, the polynomial equation ( 6 ) can be reduced to the simple relation

5

E

-

-30 -25 I ~ U C I - I rn01.1-l) ,

Figure 5. Plots of Eo’ vs. In [Cl-] for the Sb(III)/Sb couple in acidic melts: a, from potentiometry; b, from voltammograms of Sb(II1) solutions on RGCDE obtained from a reverse scan (see text).

2

e-

I

1

-35

0

5

Eo’= Eo1- ( R T / n F ) In 0,

- ( p R T / n F ) In ([L-]/Ce)

(8)

E, V

VI

AIRllllli

M solution of Sb(II1) Figure 4. Voltammograms of a 19.5 X in 1.09:l melt, at RGCDE. Rotation rate/rpm: a, 600; b, 1200;c, 1800;d, 2400; e, 3000; f, 3600. potentiometry, and a plot of this Sb(III)/Sb potential vs. In [Cl-] is shown as curve b in Figure 5 . The slope of this line is 19.8 mV, not far from the theoretical slope for p = 2; 2RT/3F = 18 mV at 40 OC (see below). Under conditions where, finally, a stable Sb(III)/Sb potential could be obtained from direct potentiometry, the potential of the Sb(III)/Sb couple was measured by conventional means and as a function of the melt composition. The potential of the couple-and the following discussion refers to both the direct potentiometry and the “zero-current crossing potential” measurement referred to abovedepends on the binding ligand concentration according to the relation Eq = Eo1+ ( R T / n F ) In (C,/Ce) ( R T / n F ) In (1 + p,[L-] + p2[L-I2 + ...) (5)

where p is the number of ligands. The slope of the curve a in Figure 5 permits evaluation of the number of the ligands, p . The experimental slope is 18.4 mV, yielding a value o f p = 2; 2RT/3F = 18 mV at 40 OC. The chloride concentrations in these acidic melts were calculated from equilibrium 1 and a value of the equilibrium constant of 9.5 X corrections for one chloride, furnished by SbC13, were made in these calculations. Thus, these plots of the potential of the Sb(III)/Sb potential in Figure 5 appear to indicate that SbC12+ is the predominant Sb(II1) species existing in the acidic melt. Provided that a value of Eo is known, a value of flSbC12+may be calculated. In the case of calculating such stability constants in chloroaluminate melts, a value of E l o is usually taken at a fixed melt composition. Table I presents values for E O ’ , defined as per eq 7 . We arbitrarily choose Eo’ for the 1.1:1 melt (which can be interpolated from the data presented to be 0.83 V), set it equal to Eo1,and extrapolate the data of curve 1, Figure 5 , to In [Cl-] = 0, to give a value of what is termed in Table I of 0.38 V (which is effectively the first two terms of the right-hand side of eq 8). We can then obtain = 4.0 X 1O2I. While the calculation of this a value of PSbCI2+

1730 Inorganic Chemistry, Vol. 23, No. 12, 1984

number is straightforward, i t s meaning is not. In the first place, as pointed out above, choice of a different melt composition to fix Eolis totally arbitrary; it may, for instance, be taken in a 0.8:l melt. In the past, Eo1values have been taken in the acidic melts during attempts to obtain formation constants for chloro complexes in the basic Since such EoI values depend on the mole ratio of aluminum chloride to organic chloride, values of formation constants are also dependent on the choice of Eo I values. Far more important, however, is the fact that a calculation involving a [Cl-] value based on application of the solvent equilibrium constant, such as carried out here, depends on the specific value of that constant (taken here as 9.5 X 10-13) to construct a plot such as in Figure 5; the values for In [Cl-] are found from the mole ratio of melt employed and the equilibrium constant. A somewhat different value of the solvent constant will simply shift the In [Cl-] scale along the axis. It will not change the slope-hence the value for the number of ligands involved in the coordination with Sb(II1) in the acid melt-but will drastically alter the value of the derived DsbCl2+. This problem is far less serious in determining chloro-complex constants in a basic melt, since in such a case the concentration of C1- is derived from stoichiometric considerations. We point this out because we have recently obtained evidence pointing to the fact that reported equilibrium constants for the BuPyCl-AICl, systemlo and/or the related imidazolium chloride-aluminum chloride systems2*may be significantly in error.29 While we recognized the problems inherent in such measurements in our earlier work,1° in which we indicated that our measured values for the equilibrium constant for eq 1 in the BuPyC1-AICl3 melt were an upper limit, the extent of the error and its consequences do not appear to have been realized. The formation of SbC12+in these acidic melts, as indicated from the above results, is in agreement with its formation in SbC13 melts containing AlC13.14~16,18 The validity of the Stokes-Einstein relation, as was established earlier, gives support for the formation of one Sb(II1) species across the entire acidic region and gives a value of 1.93 A for the radius of SbC12+species. This radius would appear to be rather small; data would indicate an ionic radius of 3-5 A might be expected, depending on the geometry of the SbC12+species in solution.30 Similar calculations based on the finding of the constancy of the qD product likewise appear to yield values for r which are too smalL6 A modification, appearing to be a better approximation when it is considered that the radius of a moving particle may be similar to that of the media of viscosity q in which it moves, suggests that the 6 in the Stokes-Einstein equation be replaced by a 4. Under those conditions, the values calculated for r are increased by Since the constancy of the qD product in these melts appears well verified experimentally, it is not unreasonable to suggest that the use of the modified Stokes-Einstein equation, as proposed by M ~ L a u g h l i ngives , ~ ~ a better indication of the ionic radii of the diffusing species. No consideration is given to solvation, since our measurements give no information regarding that possibility. Basic Melts. Table I also lists data obtained from the cathodic wave in cyclic voltammograms (Figure 1, curve b) of solutions of SbC13 in basic melts. A negative shift in the cathodic current peak with increasing scan rate is similar to that observed in acidic melts and indicates either a slow charge-transfer reaction (irreversible behavior) or a nucleation overvoltage. The hysteresis shown on the reverse scan is typical (28) Wilkes, J. S.; Levisky, J. A,; Wilson, R.; Hussey, C. L. Inorg. Chem. 1982, 21, 1263. (29) Karpinski, Z.; Osteryoung, R. A. Inorg. Chem., in press.

(30) "Handbook of Chemistry and Physics", 60th ed.;CRC Press: Boca Raton, FL, 1980; p F214. (31) McLaughlin, E. Trans. Faraday SOC.1959, 55, 28.

Habboush and Osteryoung

E , V v e AI/AIlIII)

Figure 6. Voltammograms of a 20.4 X IO-' M solution of Sb(II1) in 0.75:l melt, at RGCDE. Rotation rate/rpm: a, 400; b, 900; c, 1600;, d, 2500; e, 3600.

a112

sec-1/2

Figure 7. Limiting current vs. w 1 I 2 in 0.75:l melt: a, 20.4 M Sb(II1); b, 10.2 X M Sb(V).

X

of that noted for nucleation overvoltage deposition processes. A negative shift in the peak potential with decreasing basicity (increasing pC1) of these solutions may be taken as an indication of a chloride coordination. The charge under the anodic stripping peak was equal to that under the cathodic reduction peak. Voltammograms of the cathodic wave at the RGCDE showed no limiting region due to the proximity of the Sb reduction to the negative limit of the melt (at -1.1 V on glassy carbon). Voltammograms on the RGCDE of a solution of SbC13in the 0.75:l melt, obtained by scanning anodically, are shown in Figure 6. As shown in Figure 7, curve a, a plot of the limiting anodic currents (i,,) at potentials on the plateau, vs. uI/2was linear and passed through the origin, indicating that the reaction is convective diffusion controlled at the limiting plateau. Similar plots of the currents at potentials on the ascending portion of the wave, however, were not linear, in-

Inorganic Chemistry, Vol. 23, No. 12, I984 1731

Electrochemical Studies of Sb(II1) and Sb(V)

I

I

-2

I

I

-1

I

I

0

In(lct-1, mot. 1-1)

Figure 8. Potentiometric data for the Sb(III)/Sb couple in basic melts as a function of In [Cl-1. dicating some kinetic complications. The oxidation state of the species present in solutions of SbC13 in the basic melts was established by coulometry in a manner similar to that described fur the acidic melts; coulometric deposition of S b was carried out at -0.9 V, and at several stages the limiting anodic currents at the plateau of the anodic wave (Figure 6 ) were measured. The results of this experiment indicated the presence of Sb(II1) in solution; Q, found by extrapolation for complete deposition was 15.4 C compared with that calculated for a complete reduction of Sb(III), 15.3 C. Potentiometry of the Sb(III)/Sb couple, in different melt compositions, was performed to study the nature of the complex Sb(II1) species present in solution. A plot of E O ’ , calculated from the potentiometric data as previously described for the acidic melts, vs. In [Cl-] was linear (Figure 8 ) . The experimental slope of the plot in Figure 8 at 36.7 mV, yielding a value for the number of ligands of p = 4; 4 R T / 3 F = 36 mV at 40 OC. The chloride concentration in these basic melts was calculated from the amount of BuPyCl in excess of the 1:l molar ratio of A1Cl3:BuPyCl. Provided that a value of E o l is known, a value of a, may be calculated from eq 8. As discussed above, a value of Eo1 is usually taken at a defined melt composition; again, it must be realized that any melt composition could be employed to determine a value for Eol, which would result in different values for the stability constant. The 1.1:l melt has usually been chosen: assuming that the chloride concentration in this melt is negligible (Le., no complex formation). This method can be used to estimate a stability constant of the SbC14species although in the acidic melts Sb(II1) was shown to exist as SbC12+. Taking Eo’ in the 1.1:l melt as 0.83 V and setting it equal to Eo1as before, and the value for Eo’ of -0.52 V, obtained from Figure 8 by extrapolation to In [Cl-] = 0, we obtained a value of flSbcI,- = 1 X The ratio of the stability constants of SbC14- to SbC12+can also be calculated. In Table I a formal potential, is defined and its values for different melt compositions are tabulated. As expected from its definition, these values are constant, within tolerable limits, throughout the acidic region and the basic region. This potential, however, can be related to E o 1 by the relation EoT2= E o I - ( R T / n F ) In 8, (9) Therefore, it is dependent on the stability constant of the Sb(IJ1) species present in solution. Thus, the difference between its values in the basic melts and in the acidic melts can be directly related to the relative stability of SbC14- and SbC12’. From the average values obtained for in both

Figure 9. Potentiometric data for the Sb(V)/Sb(III) couple in the 0.751 melt.

acidic and basic melts, the ratio of the stability constants of SbC14-to SbC12+was calculated to be 1 X 1044 mol-2 L2.This ratio is independent of any arbitrary choice of Eo1 (provided that they are the same for both acidic and basic melts) since the latter cancels out during its calculation. It is, however, dependent on the value for the solvent dissociation constant. To establish the oxidation state of the product of the anodic oxidation of the Sb(TI1) solutions, constant-potential coulometry for the oxidation was carried out at + O S V (see Figure 6 ) . The results of this experiment showed that the oxidation state of the product is + 5 . Q, found for complete oxidation was 82.5 C compared to that calculated for complete oxidation of Sb(II1) to Sb(V) of 8 1.9 C. The results of potentiometric measurements of the Sb(V)/Sb(III) couple on an inert electrode (glassy carbon) during coulometry gave a linear plot (Figure 9 ) of the potential vs. the natural logarithm of the ratio of concentrations of Sb(V) to Sb(II1) species, as expected from the relation E , = Eo’

+ ( R T / 2 F ) In (jSb(V)]/[Sb(III)])

(10)

where Eo’ is the standard formal potential for Sb(V)/Sb(III) species. The slope of the line in Figure 9 was 14.5 mV (RT/2F is 13.5 mV). Potentiometry of the Sb(V)/Sb(III) system as a function of the melt composition was performed to obtain an indication of the difference between the number of chloride ligands bound to Sb(V) and Sb(II1). As shown in Figure 10, a plot of E O ’ , obtained from eq 10, vs. In [Cl-] was linear over the melt composition range 0.751 to 0.97:1,indicating one dominant complex for each of the Sb(II1) and Sb(V) species. Thus a polynomial equation, similar to eq 6 , for Sb(V)/Sb(III) can be reduced to the simple equation Eo’ = - ( R T / 2 F ) In CPp,/B,,> ((Pz - P , ) R T / 2 F ) In [ C l - I / P (11)

where Eo1is the formal standard potential of Sb(V)/Sb(III) in the absence of free chloride (i.e., no complex formation), PPI and a,, are the stability constants for Sb(1II) and Sb(V) complexes, and p 1 and p 2 are the number of ligands bound to Sb(II1) and Sb(V). The experimental slope of the plot in Figure 10 is 26 mV, yielding a value of (p2-pl) = 2; 2RT/2F = 27 mV. Further calculation to obtain the ratio of &,2//3pl, relative to a 1.1:l

1732 Inorganic Chemistry, Vol. 23, No. 12, 1984

g crl-' NAINB s0.75 0.514 0.79 0.425 0.82 0.370 0.83 0.355 0.87 0.310 0.91 0.275 0.92 0.270 0.93 0.260 0.97 0.240 a

DSb(III);

cm

qDSb(III),

l o + g cm S-?.

S-1

0.79 0.93 1.08 1.12 1.32 1.50 1.53 1,59 1.69

DSb(V);

qDSb(V),

S-1

5-l

IO-' cm

4.06 3.95 4.00 3.98 4.09 4.13 4.13 4.13 4.06 av 4.06 f 0.07

1.57 1.90 2.17 2.27 2.58 2.97 2.95 3.15 3.18

g cm

8.07 8.08 8.03 8.06 8.00 8.17 7.96 8.19 7.63 av 8.02 -i: 0.16

Reference 6.

I ~ ( I C I - Imol. , I-')

Figure 10. Potentiometric data for Sb(V)/Sb(III) as a function of In [Cl-1.

15.

10.

5 .

a

Y

0 .

9 I

-5

'

-10

'

I '

0.7

,

0.6

I

I

0.5 0.4 0.3 E , V vs AI/AI(III)

I

I

I

0.2

0.1

0

Figure 11. Voltammograms of a solution of Sb(II1) and Sb(V) in 0.751 melt, at RGCDE ([Sb(III)] = 10.2 X lo-' M; [Sb(V)] = 10.2 X lo-' M). Rotation rate/rpm: a, 400; b, 900; c, 1600; d, 2500; e,

3600.

melt or any acidic melt of fixed composition, was not attempted. A ratio calculated in this way will not be useful, since it is relative to the ratio of stability constants of Sb(V) and Sb(II1) complexes in the acidic melts, that is an unknown Sb(V) complex species and SbC12+. Although several chloro-complex species of Sb(II1) have been reported in the literature, i.e., SbC1;-,3* and SbC1,,15J6*33 only the last species has been specifically suggested in solutions of KCl in molten SbC13;15J6possible formation of some uncharacterized higher chloro complexes of Sb(II1) in molten mixtures of KCl and SbC13has been suggested." On the other hand, the only chloro complex of Sb(V) reported is SbC16-.34 From this and the results above, the formation of SbC1,- and SbC16- in the basic melt seems reasonable. Figure 11 shows voltammograms, obtained on the RGCDE, for a solution of Sb(II1) and Sb(V) in a 0.75:l melt. A plot (32) Szymanski, H. A.; Yelin, R.; Marabella, L.J . Chem. Phys. 1967,47, 1877. (33) Ahlijah, G . Y.; Goldstein, M. J . Chem. SOC.A 1970, 326. (34) Burgard, M.; MacCordick, J. Inorg. Nucl. Chem. Lett. 1970, 6, 599.

0

I

I

I

0 4 2 , rrcl/2

Figure 12. Plots of I'vs. u - ' / ~ for 20.4 X lo-' M Sb(1IIe in 0.75:l melt. Potentials from top to bottom: 0.38,0.39,0.40,0.42,0.44,0.46, 0.60 V.

of il,, at potentials on the limiting plateau, vs. was linear (Figure 7b), indicating that the reduction of Sb(V) to Sb(II1) is diffusion controlled on the plateau. Thus, from the Levich equation (eq 3) and il, and ilc for Sb(II1) and Sb(V), the diffusion coefficients for Sb(II1) and Sb(V) can be calculated, and the results of these calculations are tabulated in Table 111. The 3 0 products, also tabulated in Table 111, were found to be quite constant across the entire basic region for both Sb(II1) and Sb(V) species. Thus, the Stokes-Einstein relation was obeyed. Values calculated from eq 4 give 5.56 A for the Sb(II1) and 2.86 8, for the Sb(V) species. A smaller ionic radius for the Sb(V) species than for the Sb(II1) species is expected in view of the smaller ionic radius of Sb(V). This large difference between the ionic radii, however, may indicate Sb(II1) to be more solvated than the Sb(V) species, perhaps

Inorganic Chemistry, Vol. 23, No. 12, 1984 1733

Electrochemical Studies of Sb(II1) and Sb(V) Table IV. Dependence of oi and ke for the Sb(III)/Sb(V) System on Melt Composition NA/NB

a

0.75 0.75 0.80 0.91

0.26" 0.50b OSOb 0.50b

"

From rotating disk experiment. cyclic voltamm etry.

ke, cm s-l 1.4" 1.8' 3.4c 3.8c Assumed value.

From

by AlC1,. Comments above regarding the applicability of a modified form the Stokes-Einstein equation are still valid. As shown in Figure 12, plots of I-', at various potentials on the rising portion of the RGCDE waves in Figure 6 , vs. u-lt2 were linear with an intercept indicating a charge-transfer rate determining process. These plots were used to evaluate the rate constant according to

1 / i = 1 /ik

+ 1 /0.620nFAC*D2/3s1/6w1/2 (12)

where

ik = nFAkfC*

E

(13)

v,mVs-'

0.75

5 10 20 50 100 200

0.79 0.80

5 10

20 50 100 200 500 0.82 0.83 0.87 0.91

5 10 20 50 100 200

500 0.92 0.93 0.97

10" M Sb(II1) in

E, - EPl2vs. u1l2 was linear, indicating a quasi-reversible behavior. At low scan rates the value for Eh - Ep12approaches that for a reversible process. Thus, Ell, was estimated for these scans from the relation

(14)

Ep/2 =

E,/2

+ 1.109RT/2F

(15)

and the results are shown in Table V. These values of E,l2 were used to calculate E"', from the relation Eo' = E 1 / 2

+ (RT/ZF) In [BSb(V)/~Sb(III)1112(16)

Values of E'" obtained in this way were in good agreement

E,,,, and E"' for Sb(V)/Sb(III) on the Melt Composition 'Pa

NA/NB

X

0.75:l melt.

where ke is the rate constant at a formal potential EO', LY is the transfer coefficient, and n, is the number of electrons involved in the rate-determining step. Values of ke at Eo' obtained from eq 10 for n, = 1 are given in Table IV. Analyses of cyclic voltammograms, for the anodic wave and its corresponding cathodic wave in the reverse scan for three different basic melt Sb(II1) solutions, are given in Table V. The shift in the peak potentials with scan rate also indicated a kinetically controlled electron-transfer process. A plot of Table V. Dependence of E, ,E, , AE,, E,iz, Epa

o - ~ v

Figure 13. Plot of In kfvs. ( E - E O ' ) for 20.4

and k f is the charge-transfer rate constant at a given potential. A plot of In kfvs. E (Figure 13) is linear, as expected from the relation for an anodic process

k f = ke exp[-(1 - a)n,F(E - Eo')/RT]

- EO', ~

-

Epa,V 0.425 0.425 0.435 0.450 0.455 0.480

E,', V

AE,, mV

E","V

E",b V

l?','c V

0.370 0.365 0.365 0.350 0.340 0.330

55 60 70 100 115 150

0.385 0.385 0.390 0.400 0.400 0.415

40 40 45 50 55 65

0.400 0.400

0.405 0.405

0.407

0.410

0.430 0.430 0.435 0.445 0.455 0.470 0.490

0.390 0.380 0.370 0.370 0.360 0.350 0.335

40 50 65 75 95 120 155

0.395 0.395 0.400 0.400 0.405 0.410 0.415

35 35 40 45 50 60 75

0.410 0.410

0.415 0.415

0.412 0.414

0.410 0.410

0.417 0.420 0.427 0.437

0.411 0.41 2 0.41 1 0.41 1

0.439 0.442 0.46 2

0.410 0.409 0.406

0.445 0.450 0.455 0.465 0.475 0.4 85 0.510

0.400 0.400 0.395 0.385 0.380 0.365 0.350

45 50 60 80 95 120 160

V

0.415 0.415 0.415 0.420 0.425 0.425 0.435

E P l 2 , m V JY,,~,V

30 35 40 45 50 60 75

0.430 0.430

0.435 0.435

Habboush and Osteryoung

1734 Inorganic Chemistry, Vol. 23, No. 12, 1984

J/ = (DR/D~)(’-~)/~~~[DR?~U(~F/RT)]~’ (17) was estimated from the variation of E, - E, with scan rate. Data for ke, obtained in this way assuming a = 0.5,are reported in Table IV and are in reasonable agreement with that obtained from the RGCDE voltammograms. An attempt was made to draw conclusions on the mechanism of the electrode process, and models 1 and 2 were suggested. The symmetry of the voltammograms at the RGCDE model 1 (EEC mechanism) I 0.7

0.6

0.5

0.4

0-3

0.1

0.2

0

E. V vs AI/AI(III)

Figure 14. Comparison between an experimental RGCDE voltammogram (solid line), obtained for a solution of 13.7 X lV3 M Sb(II1) M Sb(V) in 0.75:l melt with a rotation rate of 1600 and 6.7 X rpm, and those calculated by using assumed parameters for and EEC = Ps,2 = 0.407 V, kes,I= cm mechanism (al = 0.5): - - -,Ps,l s-I; - PS,’ = Pb2 = 0.407 V, keq1 = lo-’ cm s-’; -,PSI = 0.500

.-,

- e

SbC16- 4- e-

SIOW

SbC162-

SbC16’- 4- eSbC163-

fast

fast

kes,l,CY,

SbCls3- Ee s.2

SbC1,

+ 2c1-

SbC14-

2c1-

-

V,Ees~2=0.314V,kes,l= lO-’cm~-’;-...-,E~~,~ =0.300V,Ee,,2 = 0.514 V, ke,,l = lo-’ cm s-l.

model 2 (EEC mechanism)

SbC163-

-

net: SbC16- 4- 2e-

-

SbC14-

+ 2c1-

(Figure 1 1) supports a slow double one-electron-transfer process (model 2); nevertheless, both models were considered. Steady-state waves calculated from various sets of parameters for both models are shown and compared with an experimental I 0.7 0.6 0-5 0.4 0.3 0.2 0.1 0 voltammogram in Figures 14 and 15. These calculations gave support for model 2 and, as expected, calculated waves for E, V V S AI/AI(III) model 1 were asymmetrical, unlike the experimental wave; the Figure 15. Comparison between an experimental RGCDE voltamonly symmetrical wave based on model 1 (Figure 14) was mogram (solid line), obtained for a solution of 13.7 X IO-, M Sb(II1) much steeper than the experimental one, and certainly the M Sb(V) in 0.75:l melt with a rotation rate of 1600 and 6.7 X value of the rate constant used to calculate this wave was 1 rpm, and those calculated by using assumed sets of parameters for an EEC mechanism ( a l = a2 = 0.5, Ee,,l = 0.200 V, Ees,2= 0.614 order of magnitude higher than the values reported in Table V). IV. = kes,2(cms-1):-~~-,0.1;---,0.02,-~-,0.01;-~~~-,0.001. with those obtained from potentiometric measurements (Table V) * The peak separations E, - EPc,reported in Table V, were used to estimate ke by using a method reported by Nicholson.35 According to this method, a parameter defined by

+,

(35) Nicholson, R. S.Anal. Chem. 1965, 37, 1351.

Acknowledgment. The aid of Dr. Tadeusz Hepel in performing the calculations on the mechanism of the electrode reaction is gratefully acknowledged. This work was supported by the Air Force Office of Scientific Research. Registry No. SbC12+,73740-59-7; SbC14-, 18443-80-6; SbCl;, 17949-89-2;AlCl,, 7446-70-0;SbCl,, 10025-91-9;Sb, 7440-36-0; C, 7440-44-0; BuPyC1, 1124-64-7.